Food Chemistry 477 (2025) 143458

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Contribution of post-harvest processing in cocoa bean: Chemometric and

metagenomic analysis in fermentation step
Dulce Velásquez-Reyes a,d , Pedro García-Alamilla b, Manuel R. Kirchmayr c,
Eugenia Lugo-Cervantes a , Anne Gschaedler c,*
a
Food Technology Department, Centro de Investigación y Asistencia en Tecnología y Diseño del Estado de Jalisco (CIATEJ), A.C., Camino Arenero 1227, 45019 El Bajío,
Zapopan, Jalisco, Mexico
b
Academic Division of Agriculture and Livestock Science, Universidad Juárez Autónoma de Tabasco, Carretera, Villahermosa-Teapa Km 25, La Huasteca, Centro,
Tabasco 86280, Mexico
c
Industrial Biotechnology Department, Centro de Investigación y Asistencia en Tecnología y Diseño del Estado de Jalisco (CIATEJ), A.C., Camino Arenero 1227, 45019
El Bajío, Zapopan, Jalisco, Mexico
d
Institute of Applied Chemistry, Universidad Veracruzana, Luis Castelazo Ayala s/n, Col. Industrial Animas, 91190 Xalapa-Enríquez, Veracruz, Mexico

A R T I C L E I N F O A B S T R A C T

Keywords: Cocoa fermentation is a heterogeneous process, exhibiting a high degree of diversity of flavor, aroma, and mi­
Post-harvest practices crobial communities. A study was conducted to examine cocoa fermentations in five municipalities of a region in
Cocoa fermentation Mexico, with the objective of associating post-harvest practices, geographic area, and chemical and microbio­
Volatile compounds
logical profiles. Through the application of high-performance DNA sequencing, the microbial diversity was
identified, and the non-volatile and volatile compounds were identified and quantified by UHPLC-RID/PDA and
HS-SPME/GC–MS, respectively. Using PCA, PLS regression and Pearson correlation, post-harvest practices,
geographical factors, microbial communities, and volatile and non-volatile compounds were made. The absence
of control in cocoa fermentation was associated to Aspergillus, Escherichia, and Bacillus, and reduced the pro­
duction of essential acids for aroma. This study provides data on the diversity of post-harvest practices and their
impact on cocoa quality.

1. Introduction In general, microbial dynamics commence with an increase in the

microbial load of yeasts during the first 48 h following the initiation of
Cocoa beans are derived from specific geographical regions that are fermentation (De Vuyst & Leroy, 2020). Yeasts, in conjunction with
conducive to their cultivation, characterized by soil types, humidity lactic acid bacteria (LAB), have pectinolytic enzymes that progressively
levels, climates, and altitudes (Fanning et al., 2023; Vandecandelaere, degrade the pectin present in the cocoa mucilage, thereby facilitating a
2010). The distinctive aroma and flavor profile of cocoa beans can be transition to an aerobic environment (Daniel et al., 2009; Papal­
attributed to the presence of standardized practices during post-harvest exandratou et al., 2011; Verce et al., 2021). Following a 48-h period,
processes, such as fermentation and drying, which are crucial in contingent upon the employed turning technique, there is the potential
ensuring the consistency of the final product (Fernández-Niño et al., for an augmentation in the microbial load of LAB or acetic acid bacteria
2021; Rodríguez-Campos et al., 2011). (AAB) may increase (Camu et al., 2008; Hamdouche et al., 2019;
The fermentation process of cocoa can vary significantly due to a Velásquez-Reyes et al., 2021). Given the technological advancements in
multitude of factors, including environmental influences, the specific the field, techniques based on high-throughput DNA sequencing have
genotype of the cocoa beans (such as Criollo, Forastero, Trinitario, gained popularity due to their ability to reveal the composition and
Nacional, among others), the applied processing methods, and the identity of microbial communities in cocoa fermentations (Agyirifo
geographical origin of the cocoa beans. These factors have been shown et al., 2019; Fernández-Niño et al., 2021; Papalexandratou et al., 2019;
to influence microbial dynamics and the profile of volatile and non- Serra et al., 2019; Verce et al., 2021).
volatile compounds (Bagnulo et al., 2023; Bortolini et al., 2016; Cocoa fermentation is a pivotal process that dictates the quality of
D’Souza et al., 2017; Hanifah et al., 2022). the cocoa bean. However, it is also a multifaceted process that frequently

* Corresponding author.

https://doi.org/10.1016/j.foodchem.2025.143458
Received 13 September 2024; Received in revised form 8 February 2025; Accepted 15 February 2025
Available online 20 February 2025
0308-8146/© 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

lacks standardized protocols, resulting in variability in the aromatic region, including Paraíso, Comalcalco, Cunduacán, Cárdenas, and Hui­
profile and flavor of the final product (Herrera-Rocha et al., 2023). manguillo. Table 1 presents the characteristics of cocoa bean samples
Consequently, several studies have postulated that conducting research obtained from eighteen spontaneous cocoa fermentations within the
within a single country would facilitate the delineation of a geographic Chontalpa region. The samples were collected during the 2021 to 2022
area in which the cocoa fermentative microbiota is present, thereby cocoa harvest seasons (November to November).
contributing to the potential labeling and traceability of the quality of The fermentation processes were carried out in accordance with the
the cocoa of origin. established local practices for each sampled location, with no modifi­
The presence of microorganisms in spontaneous cocoa fermentations cations was made to obtain the identification of the real microbiota of
in various cocoa-producing countries has been identified through the spontaneous cocoa fermentations carried out in each municipality.
Illumina-based sequencing (de Almeida et al., 2019; Ouattara & Niamké, Sampling was conducted within each fermentation at three distinct
2021; Papalexandratou et al., 2019). Consequently, these microorgan­ stages (at the beginning, intermediate (after turning the cocoa beans),
isms have been linked to the geographical origins of cocoa and the di­ and final). A total of 500 g of cocoa beans were aseptically collected
versity and abundance of specific microorganisms, such as in the from various locations within the boxes and at a depth of 30 cm beneath
Amazon region, where a higher relative abundance of AAB was observed the surface of the cocoa mass. Subsequently, the cocoa beans were stored
compared to expected levels (Serra et al., 2019). Furthermore, the frozen at − 20 ◦ C until use.
implementation of such studies facilitates the comparison of identified The samples were coded with the initial two letters of the munici­
microorganisms across continents. For instance, Asia, America, and Af­ pality of origin (e.g., Pa for Paraíso, Co for Comalcalco, Cu for
rica share more than ten genera of yeasts and bacteria, which are the Cunduacán, Ca for Cárdenas, and Hu for Huimanguillo) and assigned a
most prevalent genera in cocoa fermentations worldwide (Almeida number to differentiate them within the same municipality, but at a
et al., 2019; Ouattara & Niamké, 2021; Papalexandratou et al., 2019; different location of fermentation (farm or locality). Finally, the study
Serra et al., 2019). Despite the existence of numerous studies that have noted whether the sample corresponded to the beginning, intermediate,
associated microorganisms with post-harvest processes in cocoa, these or end of cocoa fermentation.
investigations are typically constrained to specific parameters, such as
turning, without assessing the uniformity of these practices across the 2.2.1. Geographical and post-harvest factors
region or the geographical origin of the cocoa (Hamdouche et al., 2019; As previously indicated, the cocoa fermentation samples originate
Velásquez-Reyes et al., 2021). In addition, metagenomic and chemo­ from various municipalities within the Chontalpa region. Consequently,
metric analyses have identified microorganisms associated with volatile geographical factors such as municipality, altitude, temperature, and
and non-volatile compounds; however, these studies generally focus on annual average environmental humidity were evaluated (Table 2). To
a single site without considering microbial variations between different assess the local practices of cocoa fermentation, the dimensions of the
locations (Chang et al., 2024; Lima et al., 2021). fermentation vessels were recorded. In this study, fermentation vessels
The Chontalpa region is one of the regions with the highest cocoa with a capacity of less than 100 kg were designated as “Small boxes,”
production in Mexico and includes five municipalities: Huimanguillo, while those with a capacity exceeding 100 kg were classified as “Big
Cárdenas, Comalcalco, Paraíso, and Cunduacán. boxes”. The identification of cocoa fermentation practices was facili­
The objective of this study was to describe and interact with of the tated by the existence of records pertaining to the date of entry of raw
different factors that influence multiple fermentations of cocoa beans cocoa, the duration of cocoa fermentation, and the performance of in­
delimited by a region. The study factors were geographic factors such as ternal (within the box) and external (ambient) temperature-cutting tests
altitude, average annual precipitation and ambient temperature, and on cocoa beans. In the event that the sampled location possessed the
post-harvest practice factors such as established protocols, size of the aforementioned data, the samples were designated as “Established
fermentation boxes and whether the cocoa comes from a single origin or Protocols” due to their adherence to a structured and systematic
multiple origins. The present study addresses the dearth of information approach to the cocoa fermentation process. Conversely, if the sampling
with a multidisciplinary approach to obtain data on specifications for locations lacked such documented protocols, the samples were desig­
the creation and scientific support of regulations for postharvest prac­ nated as “No established protocols”. Furthermore, the geographical
tices such as cocoa fermentation, providing benefits to the farmer and origin of the raw cocoa was assessed, categorizing it as either single-
the consumer. origin or of diverse origins, based on its proximity to the fermentation
site. Photographic documentation of all cocoa fermentations is included
2. Material & methods in Fig. S1.

2.1. Chemicals and standards 2.3. High-throughput DNA sequencing and data analysis

HPLC-grade methanol, sulfuric acid and trifluoroacetic acid with The extraction of the total microbial DNA associated with beans was
purities higher than 98 % were obtained from Sigma–Aldrich (Stein­ carried out in accordance with the procedure reported by Camu et al.
heim, Germany). The standards for D-glucose, D-fructose, theobromine, (2008). From each sample, 20 g of cocoa with pulp were weighed in a
and caffeine (>98 % purity) were supplied by Sigma-Aldrich (Steinheim, sterile plastic bag, and 140 mL of 0.85 % (p/v) saline solution was
Germany). Analytical-grade lactic, citric, tartaric, and succinic acids added. The mixture was then homogenized for 10 min. The mixture was
were obtained from Sigma–Aldrich or Fluka (Buchs, Switzerland). subsequently filtered using sterile Whatman No. 4 paper. The filtrate
The following compounds were utilized as standards: such as 2-hep­ was subsequently stored in 50-mL conical tubes and subjected to
tanone, benzaldehyde, decanoic acid-ethyl ester, 2-pentanol, nonanoic centrifugation at 4000 rpm for 20 min at 4 ◦ C. The upper layer, or
acid, nonanal, 3-methyl-1-butanol, 2-methyl-propanal, acetophenone, “supernatant,” was then discarded, and the resulting pellet was collected
2-heptanol, octanoic acid, benzyl alcohol, acetic acid, isoamyl acetate, in a 1-ml microtube. This pellet was subsequently subjected to a second
2-heptanol, octanoic acid-ethyl ester, acetic acid-2-phenylethylester, centrifugation at 10,000 rpm for 5 min. The upper layer, the “superna­
acetic acid-2-methylpropylester and 3-methylbutanal were obtained tant,” was then removed, and the lower layer, the “pellet,” was stored in
from Sigma–Aldrich. a frozen state at − 20 ◦ C.
The Dneasy PowerSoil Prokit (Qiagen, Germany) was then utilized
2.2. Sample collection for the extraction and purification of DNA. The process was carried out
in strict accordance with the manufacturer’s instructions. At the
The samples were collected in five municipalities in the Chontalpa conclusion of the purification process, the quality of the DNA was

2
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

Table 1
The information of the cocoa fermentation samples obtained from different localities of the Chontalpa region.
No. Code Municipality Locality Coordinates Altitude* Sampled fermentation stage Cacao type Amount of
fermentative cocoa
Beginning Intermediate Final
mass (Kg)

Pa- Moctezuma 3 ◦
18 22′07.5”N
◦
1 11 X X X Criollo <100
(1) section 93◦ 14′04.6”W
Pa- Moctezuma 3◦ 18◦ 22′07.5”N
2 Paraíso 11 X X X Criollo <100
(2) section 93◦ 14′04.6”W
Pa- Moctezuma 3◦ 18◦ 22′07.5”N
3 11 X Criollo <100
(3) section 93◦ 14′04.6”W
Co- 18◦ 14′41.2”N Mixed (Criollo
4 Aldama 9.23 X X X >1000
(1) 93◦ 20′43.9”W and Forastero)
Co- 18◦ 14′41.2”N Mixed (Criollo
5 Aldama 9.23 X X X >1000
(2) 93◦ 20′43.9”W and Forastero)
Co- 18◦ 14′41.2”N Mixed (Criollo
6 Aldama 9.23 X X X >1000
(3) 93◦ 20′43.9”W and Forastero)
Co- 18◦ 14′44.6”N Mixed (Criollo
7 Aldama 11 X >10
(4) 93◦ 21′14.2”W and Forastero)
Comalcalco
Co- 18◦ 10′58.3”N Mixed (Criollo
8 Comalcalco 20.3 X X X >100
(5) 93◦ 14′26.6”W and Forastero)
Co- 18◦ 14′30.3”N Mixed (Criollo
9 Comalcalco 16.6 X X X >1000
(6) 93◦ 12′36.0”W and Forastero)
Co- 18◦ 18′36.9”N Mixed (Criollo
10 El Zapotal 14 X X >300
(7) 93◦ 15′38.8”W and Forastero)
Co- 18◦ 18′36.9”N Mixed (Criollo
11 El Zapotal 14 X >300
(8) 93◦ 15′38.8”W and Forastero)
Cu- 18◦ 07′48.1”N Mixed (Criollo
12 La Piedra 14.9 X X X >1000
(1) 93◦ 12′17.0”W and Forastero)
Cu- 18◦ 08′00.3”N
13 Cunduacán La Piedra 14.3 X X X Criollo <15
(2) 93◦ 12′59.7”W
Cu- 18◦ 07′48.1”N Mixed (Criollo
14 La Piedra 14.9 X >1000
(3) 93◦ 12′17.0”W and Forastero)
Ca- Poblado C-16 General 18◦ 06′59.3”N
15 15.8 X X X Forastero >1000
(1) Emiliano Zapata 93◦ 29′11.9”W
Cárdenas Poblado C-28
Ca- 18◦ 01′04.3”N Mixed (Criollo
16 Coronel Gregorio 22 X X X <100
(2) 93◦ 29′50.0”W and Forastero)
Méndez
Hu- 17◦ 37′07.5”N Mixed (Trinitario
17 Guerrero 43.7 X X X <50
(1) 93◦ 27′16.5”W and Forastero)
Hu- 17◦ 37′07.5”N Mixed (Trinitario
18 Huimanguillo Guerrero 43.7 X X X <300
(2) 93◦ 27′16.5”W and Forastero)
Hu-
19 Huimanguillo 42 X X X Criollo <100
(3)
*
Meters above sea level.

verified by agarose gel electrophoresis at a concentration of 1 % (w/v) and the SILVA database (version 138.1) for bacteria, based on percent
and a measurement of the 260/280 and 260/230 ratios (Nanodrop, identity (>90 %) (Coria-Hinojosa et al., 2024).
ThermoScientific, USA). Subsequently, 20 μL aliquots were prepared
and stored at − 20 ◦ C for subsequent processing. Subsequent analysis of 2.4. Sugar and organic acids
all DNA samples was conducted using the Illumina MiSeq platform,
employing a PE 2 × 250 bp configuration (Illumina, CA, USA), at The extraction and analysis of sugars and organic acids were carried
Novogene Company (Beijing, China). out in accordance with the method described by Velásquez-Reyes et al.
(2023). For the extraction of acids and sugars, 5 g of samples from the
2.3.1. Library preparation and sequencing initial, intermediate, and final times of cocoa fermentations were
Libraries were prepared and sequenced according to the protocol weighed, mixed with 20 mL of hot deionized water (75 ◦ C), and vortexed
established by Novogene Genomics (Beijing, China), and paired-end for 1 min. The mixture was subsequently filtered using Whatman #4
reads were subsequently generated. For the study of the fungal com­ filter paper. The filtrate obtained was then passed through a 0.22 μm
munity, the ITS1-5F intergenic region was amplified using the primers nylon membrane (Millipore) and deposited in amber vials.
ITS5-1737F (GGAAGTAAAAGTCGTAACAAGG) and ITS2-2043R (GCT The separation and quantification of sugars and non-volatile acids
GCGTTCTTCATCGATGC). For the study of the bacterial community, the from the samples were carried out by means of ultra-high performance
V3–4 region of the 16S rDNA was amplified using the primers 357wF liquid chromatography (UHPLC) (Waters, Milford, MA, USA).
(CCTACGGGNGGCWGCAG) and 806R-(GACTACHVGGGTWTCTAAT). The UHPLC analysis was performed with PDA 2998 and RID 2414
detectors (Acquity Arc, Waters, Milford, MA, USA). The separation
2.3.2. Bioinformatic analysis process was executed by employing an ultrapure water mobile phase
Subsequently, an analysis was conducted using the CLC Genomics that was acidified with 0.5 mM H2SO4, resulting in an isocratic flow rate
Workbench v23.0.4 (QIAGEN). The final paired reads were then merged of 0.6 mL/min. The temperature was maintained at 20 ◦ C for the sample
and trimmed to the same length. Prior to the sequence analysis, se­ and 60 ◦ C for the column (Aminex® HPX-87H (Bio-Rad) column, par­
quences that matched chloroplasts and mitochondria were eliminated. ticle size 9 μm, 7.8 × 300 mm). The compounds analyzed by an RI de­
Subsequent quality filter steps were implemented to exclude short reads, tector were glucose and fructose, while citric, malic, succinic, tartaric,
sequences with low-quality scores, and chimeras. Taxonomic classifi­ and lactic acids were analyzed by a UV detector with a wavelength of
cation was determined using the UNITE database (version 7.2) for yeast 210 nm. Quantification was performed with a calibration curve for each

3
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

Table 2
Characteristics of the place of origin and local post-harvest practices of the Cacao Grijalva designation of origin.
No. Code Municipality Characteristics of the place of origin Local post-harvest practices

Average annual Average annual Altitude*** Protocols Origin of raw Cocoa Fermentation Box
temperature* precipitation** material Size

Cocoa of single
1 Pa-(1) 26 1303 11 Established Small
origin
Cocoa of single
2 Pa-(2) Paraíso 26 1303 11 Established Small
origin
Cocoa of single
3 Pa-(3) 26 1303 11 Established Small
origin

Co- Cocoa of different

4 26.4 1847 9.23 Established Big
(1) origins
Co- Cocoa of different
5 26.4 1847 9.23 Established Big
(2) origins
Co- Cocoa of different
6 26.4 1847 9.23 Established Big
(3) origins
Co- No Cocoa of different
7 26.4 1847 11 Small
(4) established origins
Comalcalco
Co- No Cocoa of single
8 26.4 1847 20.3 Big
(5) established origin
Co- Cocoa of different
9 26.4 1847 16.6 Established Big
(6) origins
Co- Cocoa of different
10 26.4 1847 14 Established Big
(7) origins
Co- Cocoa of different
11 26.4 1847 14 Established Big
(8) origins

Cu- No Cocoa of different

12 26.1 1947 14.9 Big
(1) established origins
Cu- No Cocoa of different
13 Cunduacán 26.1 1947 14.3 Small
(2) established origins
Cu- Cocoa of single
14 26.1 1947 14.9 Established Big
(3) origin
Ca- Cocoa of different
15 26 616.6 15.8 Established Big
(1) origins
Cárdenas
Ca- No Cocoa of single
16 26 616.6 22 Small
(2) established origin

Hu- No Cocoa of single

17 26.2 2290 43.7 Small
(1) established origin
Hu- No Cocoa of single
18 Huimanguillo 26.2 2290 43.7 Big
(2) established origin
Hu- No Cocoa of single
19 26.2 2290 42 Small
(3) established origin
*
Environmental temperature ( C). ◦
**
mm.
***
Meters above sea level.

analyte (Table S1). by liquid chromatography (UHPLC) (Acquity Arc, Waters, Milford, MA,
USA). The column utilized was a C18, 2.7 μm, 4.6 × 150 mm (Cortecs®,
2.5. Alkaloids Waters, Milford, MA, USA). The mobile phase consisted of water (Sol­
vent A) and methanol (Solvent B), with a flow rate of 0.8 mL/min. The
The extraction and analysis of alkaloids were carried out in accor­ oven was set at 30 ◦ C, and the volume of the injected sample was 10 μL.
dance with the method described by Velásquez-Reyes et al. (2023). The The gradient conditions commenced with 75 % A for a duration of 4 min,
samples of cocoa fermentations underwent a dual degreasing process followed by 80 % A for 6 min and 75 % A for 8 min. Theobromine and
with 10 mL of n-hexane (Sigma–Aldrich, Germany) at 200 rpm and 30 ◦ C caffeine were detected and quantified at 280 nm (PDA 2998, Waters). It
for 10 min, followed by centrifugation to yield a sediment from the is imperative to note that all compound analyses were carried out in
samples (Hernández-Hernández et al., 2016). The degreased samples duplicate. Quantification was performed with a calibration curve for
were then placed in a fume hood overnight to facilitate solvent evapo­ each analyte (Table S1).
ration. Subsequently, the samples were stored at − 20 ◦ C until analysis.
For the extraction of alkaloids, 1 g of the previously defatted cocoa beans 2.6. Volatile compounds
was weighed and placed in an Erlenmeyer flask with 70 mL of hot
deionized water (75 ◦ C). The sample was subjected to heating at 100 ◦ C The solid phase microextraction technique with headspace mode
for 25 min, followed by cooling to 20 ◦ C. The mixture was subsequently (HS-SPME) was utilized in accordance with the method established by
filtered with Whatman No. 4 paper. Then, 25 μL of the filtered sample Rodríguez-Campos et al. (2011). The samples of cocoa fermentations
was added to an amber vial with 1000 μL of 70 % acidified methanol were ground with a mortar and pestle until cocoa flour was obtained,
with 0.1 % trifluoroacetic acid. The mixture was then subjected to a five- and then 2 g of the flour was placed in a 10 mL headspace vial. The
minute vortex agitation, followed by filtration through a 0.22 μm nylon volatile compounds in the extracts obtained by HS-SPME. The solid
membrane (Millipore). The separation of the alkaloids was carried out phase microextraction fiber utilized was a 50/30 μm divinylbenzene/

4
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

carboxen/polydimethylsiloxane (DVB/CAR/PDMS) (Supelco) fiber. The oxygen distribution and, consequently, microbial activity. The duration
fiber was equilibrated for 15 min at 60 ◦ C, after which it was exposed to of cocoa pod storage prior to cocoa bean extraction was also considered.
the samples of cocoa for 30 min at the same temperature. The volatile Furthermore, some fermentation processes were carried out without
compounds were identified by gas chromatography/mass spectrometry adhering to good hygiene practices, which could introduce contami­
(GC/MS) (Agilent, 6890 N) with a capillary column (60 m × 0.25 mm id nants and affect the quality of the final product. Poor hygiene practices,
x 0.25 μm film thickness) (Innowax, Agilent). The selective mass de­ such as covering fermentation boxes with plastic bags and combining
tector utilized was a quadrupole (Agilent Technologies, 5975) equipped damaged or diseased cocoa beans with the rest of the batch, were also
with an electronic impact ionization system at 70 eV and 260 ◦ C. The observed. The diversity in techniques and hygienic conditions un­
temperature of the gas chromatograph (GC) oven was programmed as derscores the necessity for the standardization of fermentation methods
follows: Initially, the oven was set at 40 ◦ C for 5 min, followed by in­ to ensure the quality and safety of fermented cocoa. In some cases,
cremental increases of 10 ◦ C every minute until it reached 200 ◦ C. At this obtaining samples from the beginning, middle, or end of the fermenta­
temperature, the oven was left for a duration of 30 min. High-purity tion process was not feasible.
helium was utilized as the carrier gas at a flow rate of 0.7 mL/min.
The injector was operated in splitless mode at 240 ◦ C for 0.5 min. The 3.2. Microbial community and diversity dynamics
acquisition of data was conducted through the implementation of full
scan mode, encompassing a mass range from 35 to 500 m/z. The iden­ 3.2.1. Fungal diversity
tification of compounds was based on any criteria: Firstly, a comparison This study analyzes fungal communities during cocoa fermentation
of the mass spectra with the NIST/EPA/NIH v2.0 d 2004 Library of Mass in several municipalities of the Chontalpa region. The analysis revealed
Spectra (Gaithersburg, MD, USA) was conducted. Secondly, a compari­ that the genera Hanseniaspora, Saccharomyces, and Pichia were pre­
son of the retention index with literature data was performed. Whenever dominant in cocoa fermentations across all municipalities (Fig. 1A),
possible, the identification was confirmed by using pure standards of the with Hanseniaspora guilliermondii being the most prevalent species
components. The concentration of the compounds was obtained by an (Fig. 1A).
external standard method, where calibration curves were built with each In Comalcalco, Saccharomyces cerevisiae was identified as the most
available standard compound dissolved in methanol (Table S2) prevalent species throughout the fermentation process, exhibiting a
(Velásquez-Reyes et al., 2023). The quantification method has been relative abundance ranging from 40 % to 90 %. Additionally, Kazach­
previously employed by other researchers (Rodríguez-Campos et al., stania humilis, Hanseniaspora uvarum, and Lichtheimia ramosa were
2012; Velásquez-Reyes et al., 2023). Aroma descriptors for the com­ detected in significant quantities. In contrast, cocoa fermentations that
pounds identified were obtained from online databases from Flavornet adhered to standard post-harvest practices exhibited a predominance of
(http://www.flavornet.org/flavornet.html), The Good Scents Company Hanseniaspora, Saccharomyces, and Kazachstania.
(http://www.thegoodscentscompany.com/indeX.html), and the litera­ In Cunduacán, a high abundance of Hanseniaspora, Saccharomyces,
ture (Hinneh et al., 2020; Magagna et al., 2017; Rodríguez-Campos Thielavipsis, and Saccharomycopsis were observed in fermentations from
et al., 2011). different places and with non-standardized post-harvest practices.
In Huimanguillo, H. guilliermondii, Pichia manshurica, Lichtheimia
2.7. Statistical analysis and chemometric tools ramosa, H. uvarum, and Pichia membranifaciens were identified.
In the context of small box conditions and standardized postharvest
The experimental data are presented as the mean and standard de­ practices, the fermentations at Paraíso exhibited a limited number of
viation (SD). Subsequently, an analysis of variance (ANOVA) was con­ yeast species.
ducted, followed by Tukey’s multiple range tests to ascertain the The Venn diagram (Fig. 2) reveals that Comalcalco did not exhibit
statistically significant disparities between the means of the concentra­ exclusive yeast genera but shared more than 100 genera with
tions of sugars, acids, alkaloids, and volatile compounds. The mean Cunduacán. In Cárdenas, five exclusive genera of filamentous fungi were
differences were considered significant at p < 0.05. The statistical detected: These included Collybia, Schizophyllum, Fonsecaea, Aphano­
analysis was conducted using XLSTAT 2023, v1.6 software (Addinsoft, phora, and Valsa. At the species level, the Comalcalco samples yielded
New York, USA). The normalized concentrations of volatile and non- 159 exclusive yeast species. The most prevalent species in the five mu­
volatile compounds were utilized as variables for the construction of nicipalities included S. cerevisiae, K. humilis, H. guilliermondii, A. flavus,
the principal component analysis (PCA). Furthermore, the relationships H. uvarum, T. delbrueckii, and P. manshurica, among others.
between the microbial community of cocoa fermentations and the
different local postharvest practices were investigated through partial 3.2.2. Bacterial diversity
least squares discriminant analysis (PLS-DA). Pearson correlations were In cocoa fermentations from the Chontalpa region, the bacteria with
performed based on the data on the quantification of volatile and non- the highest relative abundance were Acetobacter, Limosilactobacillus,
volatile compounds and cocoa fermentation practices in the Chontalpa Tatumella, Lactobacillus, Liquorilactobacillus, and Pediococcus (Fig. 1B).
region and the concentrations of volatile and non-volatile compounds The most prevalent species included Acetobacter pasteurianus, Limosi­
and their relationship with the major genera of filamentous fungi, lactobacillus fermentum, Tatumella saanichensis, Tatumella punctata, Ped­
yeasts, and bacteria from all fermentations. The integration of these iococcus acidilactici, Lactobacillus ultunensis, and Liquorilactobacillus
results into two heat maps facilitated the visualization of significant cacaonum (Fig. 1B).
relationships between the compounds. In Comalcalco, cocoa fermentations were predominantly character­
ized by the presence of A. pasteurianus, L. fermentum, P. acidilactici, and
3. Results L. ultunensis, accounting for more than 60 % of the observed
fermentations.
3.1. Sample collection Fermentations involving cocoa from disparate origins, in substantial
containers, and characterized by non-standardized post-harvest prac­
A total of 19 spontaneous cocoa fermentations were sampled at the tices exhibited the pronounced relative abundance of A. pasteurianus and
beginning, intermediate, and end of the process within the Chontalpa T. saanichensis in Cunduacán.
region (Table 1). The fermentation process was executed under diverse In Huimanguillo, A. pasteurianus was identified as the predominant
practices, as illustrated in Fig. S1, demonstrating significant variability bacteria (exceeding 80 % at the conclusion of fermentations).
in the process conditions. The handling of cocoa beans was observed to Conversely, cocoa fermentation in smaller boxes, accompanied by
involve various techniques, including turning practices, which affect standardized post-harvest practices and cocoa from a singular origin,

5
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

Fig. 1. Relative abundances of genus and species of A) yeasts and filamentous fungi and B) bacteria at the beginning, intermediate, and final times of cocoa fer­
mentations from Comalcalco, Cunduacán, Huimanguillo, Paraíso and Cárdenas.

Fig. 2. Venn diagram that shows the differences between yeasts and bacteria by genus and species of the municipalities of the Chontalpa region.

exhibited a notable prevalence of A. pasteurianus, L. fermentum, T. Fig. 2 illustrates the common and exclusive genera and species of
punctata, Citrobacter freundii, and T. ptyseos in Paraíso. bacteria for each municipality. A total of 126 shared genera were
Conversely, at the culmination of the fermentation process in identified in the municipalities of the Chontalpa region. Comalcalco
Cárdenas, a heightened prevalence of A. pasteurianus (>40 %), Liq­ exhibited the highest diversity of exclusive genera (>100 genera) and
uorilactobacillus cacaonum (40 %), L. ultunensis (20 %), and C. freundii the greatest number of exclusive species of yeasts and bacteria. This can
(25 %) was observed. be attributed to the higher number of cocoa fermentations sampled

6
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

(eight fermentations) in comparison to Paraíso (three fermentations) from Cunduacán, Comalcalco, and Huimanguillo.
and Cárdenas (two fermentations). The analysis identified more than The organic acids identified and quantified in cocoa beans included
100 species of bacteria shared among the municipalities of the Chon­ citric, succinic, tartaric, and lactic acids. Citric acid exhibited higher
talpa region, including species such as Leuconostoc pseudomesenteroides, concentrations at the initiation of cocoa fermentations and lower con­
Liquorilactobacillus nagelii, Limosilactobacillus secaliphilus, Limosilactoba­ centrations at the cessation of fermentation (8.8 and 1.9 g/Kg, respec­
cillus reuteri, Lactobacillus ultunensis, and AAB such as Acetobacter pas­ tively) (Fig. 3C). In the municipalities of Cunduacán and Huimanguillo,
teurianus, Gluconobacter oxydans, and Acetobacter lovaniensis. the samples Cu-(3), Cu-(2), Hu-(1), and Hu-(3) exhibited the highest
A series of Venn diagrams were constructed to facilitate the com­ citric acid concentrations at the onset of fermentation (Fig. 3C).
parison of studies that employed high-throughput DNA sequencing as a Conversely, succinic acid exhibited concentrations ranging from 0.8
methodology for identifying microorganisms in other countries to 0.2 g/Kg, with the highest concentrations observed at the initial and
(Fig. S3). The analysis revealed that S. cerevisiae and Bacillus clausii were intermediate stages of fermentation (Fig. 3D). The samples from the
the only species shared by Nicaragua, Cameroon, Ghana, Brazil, Ivory Huimanguillo municipality exhibited the highest succinic acid
Coast, and the Chontalpa region. Furthermore, the metagenomic anal­ concentrations.
ysis of the Chontalpa region unveiled the presence of over 25 species of Tartaric acid was quantified in a range of 1.3–0.5 g/Kg (Fig. 3E). At
microorganisms that are endemic to this region. The metagenomic study the conclusion of the cocoa fermentations, a heightened concentration
of the microbial community of the Chontalpa region showed clear dif­ of tartaric acid was observed, while at the initiation, the lowest con­
ferences in the identification and abundance of yeasts, filamentous centration was detected.
fungi, and bacteria, revealing an association with local cocoa fermen­ Lactic acid exhibited its maximal presence at the intermediate and
tation practices. final stages of cocoa fermentation, with the samples from Cunduacán
and Comalcalco registering the highest concentrations (less than 3 g/kg)
(Fig. 3F).
3.3. Sugars and organic acids

The sugars present in the cocoa beans from the fermentations that 3.4. Alkaloids
occurred in the Chontalpa region were quantified. The glucose levels
exhibited a higher concentration at the beginning of most samples and a A study was conducted to identify and quantify the presence of al­
lower concentration at the end of cocoa fermentations (9–7 and 1.9 g/ kaloids, including theobromine and caffeine, in cocoa beans during the
Kg, respectively) (Fig. 3A). The samples exhibiting the highest glucose fermentation process in the Chontalpa region.
levels at the beginning were obtained from Cunduacán, Comalcalco, and Theobromine was found to be present in higher concentrations in the
Huimanguillo. samples from Paraíso (>25 g/kg) (Fig. 3G). Conversely, the samples
A similar trend was observed in fructose, which exhibited higher from Cunduacán exhibited the lowest theobromine concentrations. The
concentrations at the beginning and lower concentrations at the end of concentration range of theobromine was found to be between 29.7 and
fermentation (15 and 2.9 g/Kg, respectively) (Fig. 3B). The samples with 12.4 g/Kg.
the highest fructose concentration at the initiation of fermentation were The presence of caffeine was detected in a range of 1.9–0.3 g/kg in

Fig. 3. Average concentrations (g/Kg) of A) glucose and fructose, B) citric acid and lactic acid, C) succinic acid and tartaric acid, D) theobromine and caffeine from
samples from Cárdenas, Cunduacán, Huimanguillo, Paraíso and Cárdenas at the beginning, intermediate and end of cocoa fermentations.

7
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

the cocoa bean samples (Fig. 3H). It is noteworthy that at the inception fermentations, the majority of the samples exhibited lower concentra­
of the fermentation process, the cocoa beans exhibited higher concen­ tions of acetic acid (<90 mg/Kg x 102). Octanoic acid was identified in
trations of caffeine. However, as the fermentation process progressed, higher concentration at the end of the cocoa fermentations from
the concentration of caffeine decreased, reaching lower levels at the Comalcalco and Cárdenas (>5 mg/Kg x102). Nonanoic acid exhibited
intermediate and final stages of cocoa fermentation. the lowest concentrations among the samples from Comalcalco.
The volatile compounds with the highest concentrations in the
alcohol group were 2-pentanol and phenyl ethyl alcohol. At the onset of
3.5. Volatile compounds the cocoa fermentations in the five municipalities, the highest concen­
tration of 2-pentanol was identified (500–90 mg/Kg x102). Phenyl ethyl
A total of twenty-four volatile compounds, including acids, alcohols, alcohol stood out at the intermediate point of cocoa fermentations from
aldehydes, esters, and ketones, were quantitatively analyzed during the Cunduacán, Cárdenas, and Huimanguillo (>290 mg/Kg x102).
fermentation process of cocoa from municipalities within the Chontalpa Benzaldehyde, a volatile compound belonging to the aldehyde
region (Table S3). group, exhibited the highest concentration. The highest concentrations
Among the acids, acetic acid exhibited the highest concentration. of these compounds were observed during the intermediate and final
The endpoints of the cocoa fermentations were those with the highest stages of cocoa fermentation from Huimanguillo and Comalcalco. The
concentrations (>2000 mg/Kg x 102). At the onset of the cocoa

Fig. 4. Correlation map generated by PLS-DA between the components, along the t1 and t2 axes, of relative abundances of a) yeasts and filamentous fungi, b)
majority bacteria from the cocoa fermentations of Comalcalco, Cárdenas, Cunduacán, Huimanguillo and Paraíso. Quantitative variables x = relative abundances of
microorganisms, dependent variables y = Established or non-established protocols, small or big box, single-origin cocoa, or cocoa from different origins.

8
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

initial stages of cocoa fermentation from Huimanguillo and Cunduacán protocols, as well as the production of acetic acid, phenyl ethyl alcohol,
exhibited the predominance of 2-methyl butanal and 3-methyl butanal. esters such as 2-pentanol-acetate and 3-methyl-acetate, and citric and
The ester group’s predominant compound was isobutyl acetate. The lactic acids. Conversely, unestablished protocols exhibited a positive
intermediate and final points exhibited the highest concentrations of correlation with glucose and fructose concentrations. These are ex­
isoamyl acetate. Acetic acid-2-methylpropyl ester demonstrated a pected to have a negative correlation because they are essential sub­
heightened presence in intermediate and final stages of cocoa fermen­ strates for the metabolism of yeast and lactic acid bacteria during
tations from Paraíso and Cárdenas. fermentation. Single-origin cocoa beans have been found to be positively
Acetophenone emerged as the ketone with the highest concentration correlated with aromatic markers such as 3-methyl-butanal, 2-methyl-
during cocoa fermentations at the initial and intermediate points. At the butanal, benzaldehyde, and 2-pentanol. Another heatmap was devel­
onset of Co-(5), the highest acetophenone concentration was recorded oped using Pearson correlations of the concentrations of volatile and
(184 mg/Kg x102). Conversely, 2-heptanone exhibited higher concen­ non-volatile compounds to observe their relationship with the genera of
trations during the initial phase of the Cunduacán fermentations. filamentous fungi, yeasts, and bacteria, which are the majority of all
fermentations (Fig. 5 B). The findings revealed a positive correlation
between the production of acetic acid and Acetobacter, as well as lactic
3.6. Partial least squares regression (PLS-DA), Heatmaps and principal acid-producing bacteria (LAB), including Pediococcus, Lactobacillus, and
component analysis (PCA) Limosilactobacillus. Additionally, a positive correlation was observed
between yeasts, such as Hanseniaspora, and 2-pentanol acetate and
Multiclass PLS-DA models were developed, simultaneously consid­ benzeneacetic acid ethyl ester. Pichia was identified as the most preva­
ering the variables production size, practices, and origin of cocoa with lent yeast strain in the production of esters, including 1-butanol, 3-
the relative abundances of the genus of yeast and filamentous fungi methyl acetate, octanoic, and decanoic acid ethyl ester.
(Fig. 4A) and bacteria (Fig. 4B) in the greatest presence of the cocoa The PCA, performed with the concentrations of volatile and
fermentations. This approach permitted an integrated exploration of the nonvolatile compounds, exhibited that the F1 and F2 components
relationships between variables and samples. Furthermore, the variable accounted for 58.89 % of the variance in the data (see Figs. 6 A, B, and
importance in projection (VIP) values and correlation graphs were uti­ C). Fig. 6 A illustrates the grouping of the samples by municipalities. The
lized to identify the most pertinent variables in class discrimination positive axis of the F1 component grouped the samples from Hui­
(Fig. S2). manguillo, while the negative axis grouped many of the samples from
Geographical and post-harvest factors, including “Big box”, “Cocoa Comalcalco. The positive axis of the F2 component grouped both the
of different origins”, and “Established protocols”, were found to be samples from Paraíso, Cunduacán, and many of the samples from
associated with both PLS-DA. A notable grouping was observed between Cárdenas. Fig. 6B illustrates the grouping of the samples according to
“No established protocols”, “Cocoa of single origin”, and “Small box”. cocoa fermentation times. The positive axis predominantly groups the
Of particular interest was the association of Aspergillus, Trichosporon, samples from the initial phase of cocoa fermentation, while the negative
and Lichtheimia with non-established protocols. Furthermore, fermen­ axis predominantly groups the final samples of component F1. Fig. 6C
tations carried out with cocoa from different origins, with established illustrates the PCA variables. The compounds most abundant at elevated
fermentation practices and carried out in big boxes, favored the growth altitudes include 2-pentanol, 3-methyl-1-butanol, and 2-ethyl-1-hexa­
of various LAB such as Lactiniplantibacillus, Limosilactobacillus, and nol, which are particularly prevalent in samples from Huimanguillo.
Lactobacillus. Conversely, samples from Comalcalco, Paraíso, and Cunduacán exhibi­
A heat map was constructed from the data obtained from Pearson ted elevated concentrations of theobromine, tartaric acid, caffeine, iso­
correlations of the quantification of volatile and non-volatile compounds amyl acetate, octanoic acid, and ethyl ester. The initial times of the
from local post-harvest practices (Fig. 5A). The heatmap revealed a samples are associated with fructose, glucose, citric acid, 3-methyl-
positive correlation between cocoa fermentations and established

Fig. 5. A) Heat map of Pearson correlations of volatile and non-volatile compounds associated with local post-harvest practices in the Chontalpa region. Correlation
values range from − 1 to 1. Values close to 1 represent the highest positive correlation (blue), while values closer to zero mean no linear trend between the variables;
values close to − 1 represent the negative correlation (red) between the variables. B) Heat map of Pearson correlations between volatile, non-volatile, and micro­
organisms (yeasts, filamentous fungi, and bacteria). Positive correlations are represented on a green scale (+1), and negative correlations on a red scale (− 1), indicate
the relationship between the variables evaluated. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of
this article.)

9
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

Fig. 6. Principal Component Analysis. A) Observations of municipalities. B) Observations of cocoa fermentation times. C) Variables of volatile and non-
volatile compounds.

butanal, and acetophenone. on a global scale. It is imperative to acknowledge that classic microbi­
ological methods, such as the utilization of culture media, are inherently
4. Discussion limited in their capacity to achieve the depth of metagenomic studies.
This underscores the significance of metagenomic approaches in eluci­
The present study sought to identify the associations between mi­ dating the true fermentative processes.
crobial communities and metabolites produced during spontaneous The PLS-DA demonstrated that in locales exhibiting characteristics of
fermentations of cocoa from different locations within the same region, local cocoa fermentation practices, “no established protocols” were
thereby highlighting the relationship of the factors in post-harvest associated with the presence of filamentous fungi such as Aspergillus,
practices that were identified in the Chontalpa region. Lichtheimia, Trichosporon, and Linnemannia, as well as bacteria Bacillus,
Table 2 presents the distribution of 19 cocoa samples across different Enterococcus, Escherichia, Klebsiella, among others (Fig. 4A). Earlier
municipalities, taking into account factors such as the size of the research has indicated that Enterobacteriaceae of the Tatumella and
fermentation boxes, the presence of established practices, and the origin Escherichia genera are associated with unsuccessful fermentations and
of the cocoa. With respect to the size of the boxes, 13 samples were the production of compounds that result in undesirable flavors (Hinneh
obtained from large boxes and six from small ones, with Comalcalco et al., 2018). Bacillus spp. has been shown to exert a deleterious effect on
being notable for its significant number of large boxes. With respect to the flavor of cocoa by producing short-chain fatty acids (C3-C5), which
fermentation practices, 10 samples are from established practices and 9 have been found to generate undesirable notes and a viscous texture. Its
are from non-established ones. Comalcalco and Paraíso are the munic­ presence in the range of 3 % to 8 % during fermentation has been linked
ipalities with the most samples of established practices. Finally, ten to the development of herbaceous flavors, underscoring the necessity to
samples are from cocoa of different origins, with Comalcalco again halt the process to avert sensory defects (Papalexandratou et al., 2019).
standing out with six samples of cocoa from different origins, while These findings underscore the significance of established protocols and
Huimanguillo and Paraíso have all their samples from a single origin. sanitation measures, as inadequate practices can facilitate the prolifer­
The analysis of the microbial diversity present in cocoa fermenta­ ation of undesirable microorganisms, such as fungi and bacteria, which
tions has been studied by various authors using high-throughput DNA generate compounds that adversely impact the sensory quality of cocoa
sequencing (Agyirifo et al., 2019; Fernández-Niño et al., 2021; Papal­ (Streule et al., 2022; Subroto et al., 2023; Viesser et al., 2021). Absent
exandratou et al., 2019; Serra et al., 2019). These studies have yielded a established protocols for cocoa fermentation, critical elements such as
comprehensive profile of the microbial community in cocoa fermenta­ origin, pod opening time, cocoa bean turning, cutting tests, and tem­
tions from diverse origins. However, they have not elucidated the as­ perature control remain unrecorded. Consequently, the tracking of
sociation of microorganisms with the specific post-harvest practices of sensory variation and microbial contamination issues becomes unfea­
each region. Fig. S3 illustrates the comparison of species identified sible. The development of regulations is similarly impeded.
through metagenomic studies in different countries with the present The “small box” was found to be associated with “Cocoa of single
study, highlighting the intricate nature of cocoa fermentation processes origin” that was harvested from the same place, and the microorganisms

10
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

identified included Schwanniomyces, Leuconostoc, Weissella, Tatumella, ethyl ester.

and Citrobacter. Ghisolfi et al. (2023) identified that variability in the Consequently, at higher altitudes, the concentrations of theobromine
size of fermentation boxes and the method used (box, pile, or jute bag) and caffeine are reduced. These alkaloids, which are associated with a
can affect the microbial diversity and organoleptic profile of cocoa. bitter taste and the sensation of astringency (Condori et al., 2024),
The present study described the diverse local practices of cacao suggest that cocoa beans from the Chontalpa region, due to their
fermentation in the Chontalpa region. Although the municipalities geographical location, may be less bitter. Furthermore, the chemical
belong to the same region, there is great variation in post-harvest composition of fermented cocoa beans, including minerals, volatile
practices between large and small productions, which affects the mi­ compounds, and flavanols, may vary due to differences in local prac­
crobial dynamics of fermentation. tices, soil management, tree age, and altitude (Febrianto & Zhu, 2022).
A comprehensive evaluation of post-harvest factors, including the Variations in factors such as origin, altitude, precipitation, and ambient
dimensions of the fermentation vessels, the provenance of the cocoa temperature should help identify the uniqueness of cocoa beans by
beans, and the established fermentation protocols, was conducted to origin.
ascertain the presence of volatile and non-volatile compounds in the In the context of this research, it is proposed to expand the analyzed
cocoa fermentations. fermentation sites to better reflect regional practices and to carry out a
The utilization of heat maps revealed a positive correlation between sensory evaluation that integrates chemical and microbiological find­
cocoa fermentations and established protocols, as well as the production ings, offering a comprehensive view of the factors that affect cocoa
of acids such as acetic, citric, and lactic acid, alcohols such as phenyl­ quality.
ethyl alcohol, and esters such as 2-pentanol acetate, 3-methyl acetate,
and octanoic acid ethyl ester. Saccharomyces and Hanseniaspora species 5. Conclusion
have been identified as the producers of phenyl ethyl alcohol, 2-penta­
nol acetate, and octanoic acid ethyl ester in cocoa fermentations The metagenomic study of cocoa fermentations has enabled the
(Díaz-Muñoz et al., 2023; de Ferreira et al., 2022; Velásquez-Reyes et al., characterization of the microbial community in five municipalities
2021). The positive correlation of acids such as acetic, citric, and lactic within the Chontalpa region. The implementation of established pro­
serves as an indicator that cocoa beans are undergoing internal physi­ tocols during cocoa fermentation effectively hindered the proliferation
cochemical changes, including the death of the embryo and the ex­ of deleterious microorganisms, including Aspergillus, Escherichia, Kleb­
change of organic acids resulting from the metabolism of yeast and LAB siella, and Enterococcus. The analysis revealed a positive correlation
(De Vuyst & Leroy, 2020). between single-origin cocoa beans and aromatic markers such as 3-
A positive correlation has been observed between the reduction of methyl-butanal, 2-methyl-butanal, benzaldehyde, and 2-pentanol.
sugars, including glucose and fructose, in cocoa beans and practices such Conversely, in municipalities situated at higher altitudes and experi­
as unestablished methods, single-origin cocoa, and large retail con­ encing higher average annual precipitation, there was an increase in the
tainers. The underlying factors contributing to this phenomenon war­ presence of compounds such as 2-pentanol, citric acid, and 1-hexanol-2-
rant further investigation. The invertase enzyme, which catalyzes the ethyl in cocoa beans. The study identified that the absence of established
hydrolysis of sucrose into glucose and fructose, can be of microbial protocols renders microbial diversity susceptible to undesirable micro­
origin (yeasts) or endogenous (pulp and cocoa beans) (De Vuyst & Leroy, organisms and metabolites. The establishment of protocols within local
2020). Discrepancies in the fermentation processes may influence the cocoa fermentation practices is imperative to safeguard the identity and
proliferation of microorganisms with invertase enzymatic activity, quality of cocoa from the Chontalpa region, thereby enabling the
thereby increasing reducing sugars. However, the availability of sucrose, enhancement of its unique and distinctive characteristics.
influenced by factors such as storage conditions and the geographical
origin of cocoa pods, plays a pivotal role in this process (Hamdouche CRediT authorship contribution statement
et al., 2019; Hinneh et al., 2018). Conversely, the increase in fermen­
tative mass in wooden boxes leads to a decrease in oxygen transfer, Dulce Velásquez-Reyes: Writing – original draft, Investigation.
creating an anaerobic environment that favors invertase enzyme activity Pedro García-Alamilla: Supervision. Manuel R. Kirchmayr: Writing –
(Santander Muñoz et al., 2020). review & editing, Supervision. Eugenia Lugo-Cervantes: Supervision,
Furthermore, it has been determined that the implementation of Conceptualization. Anne Gschaedler: Writing – review & editing,
fermentation practices, in conjunction with the utilization of cocoa from Funding acquisition, Conceptualization.
diverse origins and the employment of large boxes, fosters the devel­
opment of esters such as octanoic acid, ethyl ester, and decanoic acid, Declaration of competing interest
ethyl ester. These esters are recognized as markers of a fruity aroma
(Coria-Hinojosa et al., 2024). The findings underscore the necessity of The authors declare that they have no known competing financial
establishing protocols within local cocoa fermentation practices to interests or personal relationships that could have appeared to influence
ensure the consistency and quality of the final product. the work reported in this paper.
In the present study, geographic factors such as origin, temperature,
precipitation, and altitude were found to play a significant role in the Acknowledgment
differentiation of volatile and non-volatile compounds of the cocoa
fermentations evaluated. The PLS-DA and the PCA identified that the Acknowledgments to the CONHACYT doctoral scholarship corre­
compounds of 2-pentanol, citric acid, and 2-ethyl-1-hexanol are char­ sponding to CVU 880791. Part of the study was funded by the
acteristic of the fermentations originating from Huimanguillo and, CONACYT-Government of the State of Tabasco fund (project FOMIX
concomitantly, associated with elevated altitudes. This observation number TAB-2018-01-01-84312).
aligns with previous findings, where 2-pentanol was reported as a
distinctive metabolite in cocoa fermentations conducted at elevated al­ Appendix A. Supplementary data
titudes in Peru (Balcázar-Zumaeta et al., 2023). In a subsequent study,
Hanifah et al. (2022) posited that the elevated temperatures character­ Supplementary data to this article can be found online at https://doi.
istic of high altitudes, attributable to factors such as shade, sunlight, and org/10.1016/j.foodchem.2025.143458.
wind, may exert a significant influence on the metabolite profile of
cocoa beans. The high annual average temperature in Comalcalco has
been linked to the presence of phenyl ethyl alcohol and octanoic acid

11
D. Velásquez-Reyes et al. Food Chemistry 477 (2025) 143458

Data availability International, 119(March 2018), 477–491. https://doi.org/10.1016/j.

foodres.2019.01.001
Hanifah, A., Firmanto, H., Putri, S. P., & Fukusaki, E. (2022). Unique metabolite profiles
Data will be made available on request. of Indonesian cocoa beans from different origins and their correlation with
temperature. Journal of Bioscience and Bioengineering. https://doi.org/10.1016/j.
References jbiosc.2022.05.001
Herrera-Rocha, F., Fernández-Niño, M., Cala, M. P., Duitama, J., & Barrios, A. F. G.
(2023). Omics approaches to understand cocoa processing and chocolate flavor
Agyirifo, D. S., Wamalwa, M., Otwe, E. P., Galyuon, I., Runo, S., Takrama, J., & development: A review. In , Vol. 165. Food research international. Elsevier Ltd..
Ngeranwa, J. (2019). Metagenomics analysis of cocoa bean fermentation https://doi.org/10.1016/j.foodres.2023.112555
microbiome identifying species diversity and putative functional capabilities. Hinneh, M., Abotsi, E. E., Van de Walle, D., Tzompa-Sosa, D. A., De Winne, A.,
Heliyon, 5(7), Article e02170. https://doi.org/10.1016/j.heliyon.2019.e02170 Simonis, J., … Dewettinck, K. (2020). Pod storage with roasting: A tool to
de Almeida, S. D. F. O., Silva, L. R. C., Celso, G., Junior, A. C., Oliveira, G., Helena, S., … diversifying the flavor profiles of dark chocolates produced from ‘bulk’ cocoa beans?
Lopes, A. S. (2019). Diversity of yeasts during fermentation of cocoa from two sites (part II: Quality and sensory profiling of chocolates). Food Research International, 132
in the Brazilian Amazon. Acta Amazonica, 49(1), 64–70. (June 2019), Article 109116. https://doi.org/10.1016/j.foodres.2020.109116
Bagnulo, E., Scavarda, C., Bortolini, C., Cordero, C., Bicchi, C., & Liberto, E. (2023). Hinneh, M., Semanhyia, E., Van de Walle, D., De Winne, A., Tzompa-Sosa, D. A.,
Cocoa quality: Chemical relationship of cocoa beans and liquors in origin Scalone, G. L. L., … Dewettinck, K. (2018). Assessing the influence of pod storage on
identitation. Food Research International, 113199. https://doi.org/10.1016/j. sugar and free amino acid profiles and the implications on some Maillard reaction
foodres.2023.113199 related flavor volatiles in Forastero cocoa beans. Food Research International..
Balcázar-Zumaeta, C. R., Pajuelo-Muñoz, A. J., Trigoso-Rojas, D. F., Iliquin-Chavez, A. F., https://doi.org/10.1016/j.foodres.2018.05.064
Fernández-Romero, E., Yoplac, I., … Castro-Alayo, E. M. (2023). Reduction in the Lima, C. O., Vaz, A. B. M., De Castro, G. M., Lobo, F., Solar, R., Rodrigues, C., … Góes-
cocoa spontaneous and starter culture fermentation time based on the antioxidant Neto, A. (2021). Integrating microbial metagenomics and physicochemical
profile characterization. Foods, 12(17), 3291. https://doi.org/10.3390/ parameters and a new perspective on starter culture for fine cocoa fermentation.
foods12173291 Food Microbiology, 93(June 2020), Article 103608. https://doi.org/10.1016/j.
Bortolini, C., Patrone, V., Puglisi, E., & Morelli, L. (2016). Detailed analyses of the fm.2020.103608
bacterial populations in processed cocoa beans of different geographic origin, subject Magagna, F., Guglielmetti, A., Liberto, E., Reichenbach, S. E., Allegrucci, E., Gobino, G.,
to varied fermentation conditions. International Journal of Food Microbiology.. … Cordero, C. (2017). Comprehensive chemical fingerprinting of high-quality cocoa
https://doi.org/10.1016/j.ijfoodmicro.2016.07.004 at early stages of processing: Effectiveness of combined untargeted and targeted
Camu, N., González, A., De Winter, T., Van Schoor, A., De Bruyne, K., Vandamme, P., … approaches for classification and discrimination. Journal of Agricultural and Food
De Vuyst, L. (2008). Influence of turning and environmental contamination on the Chemistry, 65(30), 6329–6341. https://doi.org/10.1021/acs.jafc.7b02167
dynamics of populations of lactic acid and acetic acid Bacteria involved in Ouattara, H. G., & Niamké, S. L. (2021). Mapping the functional and strain diversity of
spontaneous cocoa bean heap fermentation in Ghana. Applied and Environmental the main microbiota involved in cocoa fermentation from cote d’Ivoire. Food
Microbiology, 74(1), 86–98. https://doi.org/10.1128/AEM.01512-07 Microbiology, 98(June 2020). https://doi.org/10.1016/j.fm.2021.103767
Chang, H., Gu, C., Wang, M., Chang, Z., Zhou, J., Yue, M., Chen, J., Qin, X., & Feng, Z. Papalexandratou, Z., Kaasik, K., Kauffmann, L. V., Skorstengaard, A., Bouillon, G.,
(2024). Integrating shotgun metagenomics and metabolomics to elucidate the Espensen, J. L., … Nielsen, D. S. (2019). Linking cocoa varietals and microbial
dynamics of microbial communities and metabolites in fine flavor cocoa diversity of Nicaraguan fine cocoa bean fermentations and their impact on final
fermentation in Hainan. Food Research International, 177. https://doi.org/10.1016/j. cocoa quality appreciation. International Journal of Food Microbiology, 304(May),
foodres.2023.113849 106–118. https://doi.org/10.1016/j.ijfoodmicro.2019.05.012
Condori, D., Espichan, F., Macassi, A. L. S., Carbajal, L., & Rojas, R. (2024). Study of the Papalexandratou, Z., Vrancken, G., de Bruyne, K., Vandamme, P., & de Vuyst, L. (2011).
post-harvest processes of the peruvian chuncho cocoa using multivariate and multi- Spontaneous organic cocoa bean box fermentations in Brazil are characterized by a
block analysis. Food Chemistry, 431. https://doi.org/10.1016/j. restricted species diversity of lactic acid bacteria and acetic acid bacteria. Food
foodchem.2023.137123 Microbiology, 28(7), 1326–1338. https://doi.org/10.1016/j.fm.2011.06.003
Coria-Hinojosa, L. M., Velásquez-Reyes, D., Alcázar-Valle, M., Kirchmayr, M. R., Calva- Rodríguez-Campos, J., Escalona-Buendía, H. B., Orozco-Avila, I., Lugo-Cervantes, E., &
Estrada, S., Gschaedler, A., … Lugo, E. (2024). Exploring volatile compounds and Jaramillo-Flores, M. E. (2011). Dynamics of volatile and non-volatile compounds in
microbial dynamics: Kluyveromyces marxianus and Hanseniaspora opuntiae reduce cocoa (Theobroma cacao L.) during fermentation and drying processes using
Forastero cocoa fermentation time. Food Research International, 193. https://doi.org/ principal components analysis. Food Research International, 44(1), 250–258. https://
10.1016/j.foodres.2024.114821 doi.org/10.1016/j.foodres.2010.10.028
Daniel, H. M., Vrancken, G., Takrama, J. F., Camu, N., De Vos, P., & De Vuyst, L. (2009). Santander Muñoz, M., Rodríguez Cortina, J., Vaillant, F. E., & Escobar Parra, S. (2020).
Yeast diversity of Ghanaian cocoa bean heap fermentations. FEMS Yeast Research, 9 An overview of the physical and biochemical transformation of cocoa seeds to beans
(5), 774–783. https://doi.org/10.1111/j.1567-1364.2009.00520.x and to chocolate: Flavor formation. Critical Reviews in Food Science and Nutrition, 60
De Vuyst, L., & Leroy, F. (2020). Functional role of yeasts, lactic acid bacteria and acetic (10), 1593–1613. https://doi.org/10.1080/10408398.2019.1581726
acid bacteria in cocoa fermentation processes. FEMS Microbiology Reviews, 44(4), Serra, J. L., Moura, F. G., de Pereira, G. V. M., Soccol, C. R., Rogez, H., & Darnet, S.
432–453. https://doi.org/10.1093/femsre/fuaa014 (2019). Determination of the microbial community in Amazonian cocoa bean
Díaz-Muñoz, C., Van de Voorde, D., Tuenter, E., Lemarcq, V., Van de Walle, D., Soares fermentation by Illumina-based metagenomic sequencing. Lwt, 106(July 2018),
Maio, J. P., … De Vuyst, L. (2023). An in-depth multiphasic analysis of the chocolate 229–239. https://doi.org/10.1016/j.lwt.2019.02.038
production chain, from bean to bar, demonstrates the superiority of Saccharomyces Streule, S., Freimüller Leischtfeld, S., Galler, M., & Miescher Schwenninger, S. (2022).
cerevisiae over Hanseniaspora opuntiae as functional starter culture during cocoa Monitoring of cocoa post-harvest process practices on a small-farm level at five
fermentation. Food Microbiology, 109. https://doi.org/10.1016/j.fm.2022.104115 locations in Ecuador. Heliyon, 8(6). https://doi.org/10.1016/j.heliyon.2022.e09628
D’Souza, R. N., Grimbs, S., Behrends, B., Bernaert, H., Ullrich, M. S., & Kuhnert, N. Subroto, E., Djali, M., Indiarto, R., Lembong, E., & Baiti, N. (2023). Microbiological activity
(2017). Origin-based polyphenolic fingerprinting of Theobroma cacao in affects post-harvest quality of cocoa (Theobroma cacao L.) beans. In Horticulturae (Vol.
unfermented and fermented beans. Food Research International.. https://doi.org/ 9, issue 7). Multidisciplinary digital publishing institute (MDPI). https://doi.org/
10.1016/j.foodres.2017.06.007 10.3390/horticulturae9070805
Fanning, E., Eyres, G., Frew, R., & Kebede, B. (2023). Linking cocoa quality attributes to Vandecandelaere, E. (2010). Geographic origin and identification labels: Associating
its origin using geographical indications. In , Vol. 151. Food control. Elsevier Ltd.. food quality with location. Innovations in Food Labelling, 137–152. https://doi.org/
https://doi.org/10.1016/j.foodcont.2023.109825 10.1533/9781845697594.137
Febrianto, N. A., & Zhu, F. (2022). Composition of methylxanthines, polyphenols, key Velásquez-Reyes, D., Gschaedler, A., Kirchmayr, M., Avendaño-Arazate, C. H.,
odorant volatiles and minerals in 22 cocoa beans obtained from different geographic Rodríguez-Campos, J., de Calva-Estrada, S. J., & Lugo-Cervantes, E. (2021). Cocoa
origins. Lwt, 153(July 2021), Article 112395. https://doi.org/10.1016/j. bean turning as a method for redirecting the aroma compound profile in artisanal
lwt.2021.112395 cocoa fermentation. Heliyon, 7(8), Article e07694. https://doi.org/10.1016/j.
Fernández-Niño, M., Rodríguez-Cubillos, M. J., Herrera-Rocha, F., Anzola, J. M., Cepeda- heliyon.2021.e07694
Hernández, M. L., Aguirre Mejía, J. L., … González Barrios, A. F. (2021). Dissecting Velásquez-Reyes, D., Rodríguez-Campos, J., Avendaño-Arrazate, C., Gschaedler, A.,
industrial fermentations of fine flavour cocoa through metagenomic analysis. Alcázar-Valle, M., & Lugo-Cervantes, E. (2023). Forastero and Criollo cocoa beans,
Scientific Reports, 11(1). https://doi.org/10.1038/s41598-021-88048-3 differences on the profile of volatile and non-volatile compounds in the process from
de Ferreira, O. S., Chagas-Junior, G. C. A., Chisté, R. C., da Martins, L. H. S., de fermentation to liquor. Heliyon, 9(4), Article e15129. https://doi.org/10.1016/j.
Andrade, E. H. A., do Nascimento, L. D., & Lopes, A. S. (2022). Saccharomyces heliyon.2023.e15129
cerevisiae and Pichia manshurica from Amazonian biome affect the parameters of Verce, M., Schoonejans, J., Hernandez Aguirre, C., Molina-Bravo, R., De Vuyst, L., &
quality and aromatic profile of fermented and dried cocoa beans. Journal of Food Weckx, S. (2021). A combined metagenomics and Metatranscriptomics approach to
Science, 87(9), 4148–4161. https://doi.org/10.1111/1750-3841.16282 unravel Costa Rican cocoa box fermentation processes reveals yet unreported
Ghisolfi, R., Bandini, F., Vaccari, F., Bellotti, G., Bortolini, C., Patrone, V., Puglisi, E., & microbial species and functionalities. Frontiers in Microbiology, 12(February), 1–24.
Morelli, L. (2023). Bacterial and fungal communities are specifically modulated by https://doi.org/10.3389/fmicb.2021.641185
the cocoa bean fermentation method. Foods, 12(10). https://doi.org/10.3390/ Viesser, J. A., de Melo Pereira, G. V., de Carvalho Neto, D. P., Favero, G. R., de
foods12102024 Carvalho, J. C., Goés-Neto, A., … Soccol, C. R. (2021). Global cocoa fermentation
Hamdouche, Y., Meile, J. C., Lebrun, M., Guehi, T., Boulanger, R., Teyssier, C., & microbiome: Revealing new taxa and microbial functions by next generation
Montet, D. (2019). Impact of turning, pod storage and fermentation time on sequencing technologies. World Journal of Microbiology and Biotechnology, 37(7).
microbial ecology and volatile composition of cocoa beans. Food Research https://doi.org/10.1007/s11274-021-03079-2

12